An iterative mathematical decision model for cloud migration: A cost and security risk approach
Mirsaeid Hosseini Shirvani,Amir Masoud Rahmani,Amir Sahafi
DOI: https://doi.org/10.1002/spe.2528
2017-09-13
Abstract:This paper presents an iterative mathematical decision model for organizations to evaluate whether to invest in establishing information technology (IT) infrastructure on‐premises or outsourcing IT services on a multicloud environment. This is because a single cloud cannot cover all types of users’ functional/nonfunctional requirements, in addition to several drawbacks such as resource limitation, vendor lock‐in, and prone to failure. On the other hand, multicloud brings several merits such as vendor lock‐in avoidance, system fault tolerance, cost reduction, and better quality of service. The biggest challenge is in selecting an optimal web service composition in the ever increasing multicloud market in which each provider has its own pricing schemes and delivers variation in the service security level. In this regard, we embed a module in the cloud broker to log service downtime and different attacks to measure the security risk. If security tenets, namely, security service level agreement, such as availability, integrity, and confidentiality for mission‐critical applications, are targeted by cybersecurity attacks, it causes disruption in business continuity, leading to financial losses or even business failure. To address this issue, our decision model extends the cost model by using the cost present value concept and the risk model by using the advanced mean failure cost concept, which are derived from the embedded module to quantify cloud competencies. Then, the cloud economic problem is transformed into a bioptimization problem, which minimizes cost and security risks simultaneously. To deal with the combinatorial problem, we extended a genetic algorithm to find a Pareto set of optimal solutions. To reach a concrete result and to illustrate the effectiveness of the decision model, we conducted different scenarios and a small‐to‐medium business IT development for a 5‐year investment as a case study. The result of different implementation shows that multicloud is a promising and reliable solution against IT on‐premises deployment.
computer science, software engineering